Online Feedback Optimization for Minimal Curtailment of Renewable Generation with Real Data

Abstract

The increasing integration of distributed energy sources into low-voltage distribution networks and of photovoltaic systems in particular poses new challenges to power grid operators. Excessive distributed generation can result in overvoltages at certain distribution feeders and may consequently lead to system damages. As an alternative to physical grid reinforcement, reactive power control (Volt/VAr control) has been shown to be an effective measure to keep the voltages within the allowed range. However, common Volt/VAr control methods are either computationally expensive or suboptimal. The aim of this work is to evaluate the performance of a recently proposed algorithm for Volt/Var control that is based on the principle of online feedback optimization (FO). For this purpose, the algorithm has been implemented on a benchmark low-voltage distribution grid which has been complemented with real data for power consumption and generation. As a baseline, all experiments are also done under the application of droop control, a local Volt/Var method that is part of many current grid codes. With respect to voltage regulation, the novel FO algorithm exhibits similar or even superior performance compared to droop control, depending on the amount of distributed generation. Moreover, the simulation results show major differences in the total actuation efforts of the two controllers. Enabled by its centralized nature, FO uses the available reactive power capabilities much more efficiently than droop control. Therefore, using FO in Volt/VAr control has the potential to considerably decrease operation costs.